Prediction of microsleeps from the EEG: A Bayesian machine-learning perspective
Speaker
Reza Shoorangiz
Institute
PhD Student, Neural Engineering Research Group, Electrical and Computer Engineering, University of Canterbury / Christchurch Neurotechnology Research Program, New Zealand Brain Research Institute
Time & Place
Fri, 23 Feb 2018 14:00:00 NZDT in Link 309 Lecture Theatre
Abstract
Microsleeps are complete arousal-related lapses of responsiveness for ~0.5–15 s, which can result in injury or death, especially in the transport sector (pilots, air-traffic controllers, truck and car drivers, etc.). The Christchurch Neurotechnology Research Programme (NeuroTech) is a world leader in lapse research in terms of the characterization and EEG-based detection of microsleeps. However, despite this achievement, the detection – and, better still, prediction – of microsleeps has proven a difficult nut to crack. This seminar presents an overview of the rationale, design, implementation, and results of a PhD project which focused on Bayesian approaches to EEG-based prediction of microsleeps.